Rethinking the Standings
Every morning when I was a kid, I would wake up and grab the sports pages to see how my team was doing. Millions of people across the country do the same — they wake up and reach for a paper to sift through something like this:
I always go straight for my division and my team, and see how many games behind they are. I then look longingly at the Wild Card standings to get an idea of how far behind they are. But most of the information gets completely ignored as I eat my breakfast. I’m guessing that millions of people across the country are doing the same.
Chris Spurlock over at Beyond the Box Score made a noble attempt to animate a division race here. But a few things didn’t sit well with me. First of all, it was out of context — you can see how your team is doing in comparison to the division leader, but not in relation to the rest of the league.
My suggestion was this:
But I still have the same problem. It shows how each team is doing in relation to the AL East, but how is the rest of the AL doing? Not to mention that it looks almost exactly like the graphs Studes used to update on The Hardball Times. Though that isn’t a bad thing, it also means that I didn’t add anything at all to the conversation.
So I got thinking even harder about it. What do we really want to see in the standings? We want to know how our team is doing. We want to know how far behind they are in the division. We want to know how they are doing in relation to the other teams in the league when it comes to the Wild Card Standings. Also, we probably want to see what direction they’re headed in. Some newspapers show the record over the last 10 games and/or the current streak (3L, or 4W, etc.). Studes created sparklines to do the same thing.
This is my attempt at making the divisional race better (the file is 6 megabytes, so please give it a little bit of time to load):
Each division is on the same graph, so we can compare across the entire AL in a glance. Whoever is furthest left in their division is in first. Each team also has a “trail” of 5 days to show what direction they are moving in. When a line pops up between two teams, it means that they have changed places in the standings.
Unlike my usual graphs, this was made entirely in Excel, and then was turned into a .gif file, so the image quality isn’t so great. Please be understanding. I also don’t handle ties too well, and don’t show which team is actually first in the Wild Card race.
I want your feedback. Does this graph work for you? Would looking at a chart like this every morning with your coffee be a good replacement for the traditional standings?
I'm an expat living in Japan since 2003, doing sales and marketing work. More of my work is available on Henkakyuu, my personal blog. Also feel free to inspire me to use twitter more often @henkakyuu
I like it. It’s cool to watch how the season played out with the full animation. My complaint on using it as the “morning coffee standings” is it’s hard to look at it and tell exactly where your team is in relation to the team(s) it’s chasing. When you look at the standings now, you can see 3 games back, or 4.5 out of the WC, etc. Stuff that you can easily translate into “What needs to happen going forward” to win this thing.
pretty freaking cool! I am more of a numbers guy than a visual guy but the movements by time ring true to me.
I think it’s a decent concept. To my eye, it would be easier to follow if the individual teams had fixed lateral positions instead of the first-place team moving to the left — the vertical position indicates that already, so it seems superfluous. Also fixing the lateral positions would reduce the model’s dependence on colour, which, of course, benefits the colour-blind and also makes it more useful in situations where it won’t print in colour (such as most newspapers).
Agree 100% with your comment about fixed lateral positions. Height alone should be enough to encode relative positions in the standings. But if you want another layer of encoding, try varying the hue in grays. That works for the colour blind and works in snapshots of the animation for newspapers also.
A very cool idea where animations are appropriate. But where they can’t work, in print for example, or where the reader doesn’t want to wait the time that it takes to view the whole thing, I think the first plot works really well. To increase the “data to ink” density enough to show six of those (i.e. all divisions), you just need to create sparkline versions of it and arrange in a grid. If the purpose is to show how your team is faring in say a newspaper, it’d be overkill to show data going back the whole season – the previous 30 days should suffice. That’d also increase the readability of the sparklines.
I’d love to show you what I mean – just need the data!
Here you go tommy:
https://spreadsheets.google.com/ccc?key=0AvhKIaAw27e_dEtiYWRRU0ZERHRZaW9FNEhjV0lWbEE&hl=en&authkey=CNOVhJsM
I think you should be able to handle it from there. It’s just the raw win totals for each team at each date in the regular season, with the last column being the Wild Card games +/- .500.
I wasn’t actually intending to animate this at first to be honest. I realize now that I only posted a moving version. The goal was to have one for each day at the season (which I could then animate at the all-star break or end of the season or whatever I suppose).
Anyway, I’d love to see what you come up with.
Well played, old top. Well played.
The graph is very nice, but it lacks the team focus that most people have. As a Jays fan, I’m always looking at how far behind I am to the division leader and also how far behind I am to the wildcard leader.
If the Jays miraculously were in front of the division, I’d want to see how much ahead I am to the nearest follower in the division or in the wildcard.
I’m glad you attempted to reinvent this concept, Josh. Both approaches have their merits, and both tell a solid story. Nicely done.
That was really cool.
This interesting if looking at the standings over a long time period, but I am not sure it works on a daily basis. Strictly for the here and now, I think the good old listings still tell the best story.
I was actually thinking the opposite. Obviously, the newspapers can’t have animation (except on their websites), so this may be a non-starter for long-term standings. But this is exactly the type of thing I could look at on a daily basis to judge where things are.
The only thing I would add might be value-keys next to each dot indicating (what the graph already indicates) the # of games back in the division… it might be hard to see if my team is 3 or 3.5 games back since the locations are so close together (or put them down at the bottom under the team abbreviations).
I’m also a fan of the idea of leaving the teams in the same order to begin the season (alphabetical seems easiest) so that I don’t have to keep track of what column my team happens to be in on any one day. With the dots clustered by division, it should still be readily visually apparent which team leads the division.
Another “graphic” thing might be to use the team logos instead of dots… in any case, superb work.
For the last chart it would be nice if there was a scale of games along the bottom, perhaps even divided into months, so you could see more exactly where in the season the animation was.
The blue bar at the bottom is a Mac excel slider, so it gives an idea. It probably isn’t clear enough though and needs improvement. I’ll try to think of a way to deal with time better.
I agree that teams should have fixed horizontal positions. I think the lines between teams that switch position are just confusing. In a daily snapshot it would be easy to annotate each team’s dot with the number of games ahead/behind in its division and/or wild card race(s). It would also be cool to see longer trails, perhaps 10 games.
I think it’s pretty cool.
Cool work Josh.
Edward Tufte suggests using sparklines in this context in his book Beautiful Evidence. You can view those pages here. Scroll down to the first comment by Tufte.
I love sparklines when talking about data density, but I don’t know if it’s best or needed in this case. And anyway, arises already did that for THT back a few years ago and I didn’t think it was as useful as it could have been (for baseball data, I tend to want more granular data for the feeling of Precision, the same way people want games behind clear on the graph). I’ll fiddle around with it and see what I can come up with.
I just got back from Tufte’s course in LA. I really like the way he presents season records with sparklines…definitely better than what we use now.
This is very cool, but how are you going to animate it in my newspaper?
p.s. all you have to do to fix the “readability” of how many games back you are is put that number right in the middle of each teams circle.
Cool idea. Definitely something to try!
Sorry, I wasn’t actually planning to animate it in the newspaper — I wasn’t really planning on animating it at all. I guess I got carried away in the moment.
Though totally transforming the sports page of the newspaper would be a fun project to do.
Ya its cool looking and is a different way to look at the league standings. But like Jack said up top, it is hard to get a sense of the exactly how many games out (or ahead) your team is. I think it is a great compliment to go above standings but I don’t think it is a replacement quite yet.
Would it be possible to make the graph interactive, such that the user could control the slider along the x axis? A way to work in some of the more traditional standings stats would be nice as well.
I love how well this graph visually represents the Baltimore nosedive at the beginning of the season.
It actually is interactive since I made it in Excel (but that’s also the cause of some deficiencies too). I can put up the spreadsheet if you’d like to see it, but it is a total mess and cheats the rules of excel in several ways.
I’d be very interested in seeing the full spreadsheet, if it wouldn’t be too much trouble to post it.
Instead of the division rankings at the bottom of the chart, change it to games back.
GB7.5 or whatever.
And change the dots to logos.
And lengthen the trail to 10 games.
But these are all nitpicks….this is a great idea.
I want to use logos but excel didn’t handle them too well (they need to be totally non-transparent squares which is a pain) so I went with circles. Hopefully for a final version I will fix that.
It’s a very cool idea. I had the same problem as a different reader: The swapping of the team dots as they changed spots in the standings was a little confusing. Both the positional swap and the color bleeding was a little tough to deal with. I think I got the most enjoyment from watching the Orioles plunge early followed by the late-season free-fall of the Mariners. Maybe those stood out more because the bad teams were farther away than the good teams from the rest of the pack.
I dislike that the horizontal position of the teams change. We can get which one is i first from the vertical position – having the horizontal position change just makes it harder to read.
Otherwise I think it’s terrific. I really like this graphic, but I do think that the team dots should only move up and down.
Do Want
Pants party?
Helpful graph. No need to keep changing the position of the teams, though. It’s obvious who is leading without having them in the extreme left position.
Surprised to see that no mention has been made of baseballrace.com
Personally I find what they do more illuminating than anything thus far discussed.
I like what is proposed here better, but that was interesting too. But it ends up with lots of lost space because most of the line is not being used at any moment.
Sorry, I should have included it. I saw it before and it probably had some impact on this. I’ll add the link when I get home tonight.
Very cool. My suggestion is to make it easier to tell how many games behind the leader you are. But great to look at for the full season.
This is very cool, Josh.
Instead of the numbers on the x-axis representing each team’s rank in the division, how about have that number be the number of games back of a playoff spot, the number of games ahead of losing that playoff spot?
Now thinking about it, maybe you can play around with it being games ahead of playoff spot/division lead/wild card, but I think that general idea works.
http://www.pennant-race.com
I think that is really cool, especially re-watching a season’s worth of standings unfold. In terms of data presentation, what if, instead of the Y axis representing winning percentage, it was “normalized” to the team leading in the wild card. That team would be at “1” (or 0 I guess and the rest of the Y axis would represent games ahead or behind). Every other team would be placed above or below the wild card team. That seems like a good way for a fan to see how “safe” his or her team is in its playoff berth or how much ground the team has to make up. It would also allow people to quickly determine how far back in the wild card their division leading team would be if it lost its division lead (assuming the wild card team had a better record than at least one division leader). Anywho, just a thought. I don’t know how easy or difficult that would be to execute.
I should have read all of the comments before I posted
Actually, it was redundant. I’m done.
Do you know about http://baseballrace.com ? Your ideas remind of it, and that place is awesome. I love the traditional standings because I like seeing the raw numbers. Your chart is cool & helpful for analyzing, but it lacks the raw W-L, PCT, GB, numbers that I thrive on seeing. Even as a kid, I loved thinking about the numbers and what my team needed to do compared to what the other teams needed to do.
So cool-i like this
This is a great first step. Works very well in many respects. Taking full advantage of the visual space is there some way to manipulate the size of a team’s dots to relate to their run differential, since that is pretty key to their place in the standings. For example a larger dot means outscoring opposition by X runs while a smaller one means X is much lower. Not sure how to deal with negative run differential. Maybe something other than a circle. Anyway, good effort and keep improving it. You have something here.
Just chiming in to say that I thought this was way cool. Now if you’ll excuse me, I have to stare at it more.
This is nice. I agree — no horizontal movement needed, “normalize” to WC leader, and put GB within circle. As far as run differential, how about having the dot be a thick black circle. A team with a dead even run differential the circle is perfectly filled with color, a positive run differential means the color extends beyond the circle and negative means the color does not entirely fill the circle.
Sorry I haven’t commented on each suggestion independently, but I read all your awesome comments and will fiddle with some variations and let you guys see what they look like. Thank you for participating.
I like the new format. It more accurately reflects how much God hates orioles fans.
This is a significant improvement over the standard format employed by mainstream sites. It would be equally, if not more, useful as a replacement for the current leader board format.
I’d love to know how you pulled this off in Excel? Years ago I was trying to create something like this for fantasy league standings, but never had much success.
I really like both. I think Fangraphs should maintain both over the course of 2011. I think that they may not be what the casual fan is looking for in their 20 seconds of baseball in the morning. But the way they tell a story is absolutely awesome (not a word I deploy lightly).
I do agree with the people suggesting fixed positions for teams, rather than moving them based on order of standings. Perhaps with a highlight for 1st. Logos instead of dots and a longer trail also make some sense.
Perhaps somebody with a knowledge of web2.0 magic could make a web app that offers those all as options?
Wow. What a terrific idea, and it looks better than I would have dreamed. The Wild Card line is a stroke of genius. Since you enjoy feedback so, I’ll throw in the two thoughts I had:
1. I wanted to see when teams were eliminated. The baseballrace site used a moving line, which on your version would appear at the bottom and move up the graph; that might work. Or maybe just a big X riding on top of an eliminated team’s dot.
1a. Come to think of it, showing when teams clinch would also be useful.
2. I wonder whether I could highlight my favorite team. Putting teams in the same horizontal position will help, but it would be so great if I could click a button to highlight, say, the Twins, and then they stand out in some way — a larger circle, “MIN” in bold text, something.
Again, this is excellent. Thanks so much for sharing it.
You just gave me about a dozen great ideas on how to do this. And 99% of them are actually doable in Excel by my estimation. And I can definitely make it super-interactive if I take the time and add it in. Very interesting idea. Awesome stuff.
i didn’t read the comments so hopefully this isn’t redundant:
The swapping of left to right confused me. I’d rather see it in Alphabetical or Geographical order with the height indicating who was in first, at least for a movie of the season. on a daily basis it might not matter.
Very cool. I agree that the leaders don;t need to always be on the left.
When the season starts, I think Fangraphs should post the current version of this chart in the corner of the homepage regularly for a while and see how people like it.
While reading this article on standings, I was struck by the three-letter acronym used by each team and the pattern used in creating it. Here’s the breakdown:
NYY: Location (2), Moniker (1) – Can’t simply use three spots for location as there is another New York club. [See CHC, CHW, LAA, LAD, NYM]
TBR: Location (2), Moniker (1) – Sometimes seen as TB, but has been updated, according to this article, after losing “Devil”. [See KCR, SDP, SFG]
ATL: Location (3) – AB for Atlanta Braves wouldn’t work as it doesn’t create a three-letter acronym, so using the first three letters of location makes the most sense. [See ARI, BAL, BOS, CIN, CLE, COL, DET, HOU, MIL, MIN, OAK, PHI, PIT, SEA, STL, TEX, WAS]
FLA: Location (2), ? (1): Why aren’t the Marlins listed as FLO?
What about creating a three-letter acronym consisting of Location (1), Moniker (2)?
CWS: Location (1), Moniker (2) – The White Sox were once listed in the standings as this three-letter acronym. Since Chicago White Sox is three words, it fits nicely according to the three-letter acronym. [See BRS for Boston Red Sox, TBJ for Toronto Blue Jays, and SLC for St. Louis Cardinals
While I am not advocating making this change, I only thought it would be interesting and could spark a (short) conversation.
[Note: LAA, LAD and NYY, NYM and CHW, CHC share the same first two letters, albeit location, as would TBR and the newly created TBJ.]
Methinks you are overthinking it a tad. 🙂 Everyone knows who the NYY, CWS, FLA, etc., are. If I saw “BRS” used as an acronym, independent of divisions or other context clues, I would have to think for fifteen seconds about who in the world it was referring to, and wehter expansion had created a new team in Bristol (Connecticut? Tennessee?).
Likewise with “TBJ”, and I’m a Jays fan. We’ll opt out and stick with TOR.
I definitely don’t think the acronyms need to change. I was just throwing an idea out there. But, no one has answered (though I doubt many have read my post) why Florida is FLA and not FLO.
I like it a lot.
Very impressive. I particularly like how you used win %, which very elegantly handles normalizing the y-axis (which would otherwise be a big problem for presenting things). A few things that I think would help:
1. Some sprite overlay/underlay which highlights the team in the lead for its division, as well as the wild card. Maybe a star or something.
2. Some other sprite that designates when a team has clinched a playoff spot and yet another sprite that designates when a team clinches its division.
3. Bigger icons. Seriously. Mine eyes strain! This could be accomplished by using icons that show the team logos, along with their initials above said icons (rather than to the side). You have tons of vertical space and very little horizontal space. It’s important to optimize your icons to spread over the vertical space more, methinks.
Right now it’s in the form:
( ) XYZ
Instead, we could be looking at:
___
( )
~~~
XYZ
Same width, much bigger icon, more readable since everything stays in the column. Not sure if it’s the easiest to do in excel, but it’s definitely something that can be done either in excel or in some more advanced plotting software (ex. python’s matplotlib, etc)
And… great. The system totally stripped my whitespace. Assume that the text in the second example formed an eliptical shape of equal width as the expression “( ) XYZ”. The point was that your dots could be just about as wide as the dot AND the XYZ, while fitting into the same amount of horizontal space, simply by moving the team initials above or below the dot.
Very cool. The division leader pseudo-acronym is dark on the bottom. The Wild card leader(s) could also be dark or just somewhat dark. You could also have a pair of standings numbers: -2.5/1.0 would mean 2.5 games back in the division and one game ahead in the wild card.
Thanks!
I have been poring over Excel for the past several hours trying to do it the way you guys want. I thought it would be simple, but:
a) I had totally screwed up last time with double-headers
b) Elimination is easy to figure out, clinching — not so much
c) You can’t go wrong with c)
This is taking me longer than projected, but I am working on it. Sorry for the delay.
It would work even better for a newspaper covering Quidditch